摘要
This study deals with the design and development process of computerized adaptive testing module integrated into UZWEBMAT. UZWEBMAT is an expert system supported adaptive and intelligent individualized e-learning environment. It was devoted to teach probability subject. A question bank was created in the study in the first place. This question bank was administered to learners in the form of work sheets. These tests were analyzed according to Item Response Theory. Model-data adaptivity was tested at the end of the analyses. It was deduced that data can be adapted to 3PL. Items non-compliant with 3PL were extracted from the question bank. At the end of these operations, final question bank containing 752 questions in total was obtained. Learners taught with UZWEBMAT took end-of-subject tests for permutation, combination, binomial expansion and probability containing 20, 20, 15 and 20 questions respectively. Besides, learners completing all subjects took end-of unit test containing 30 questions. Items in these tests were dynamically selected from the question bank pursuant to adaptive assessment. While test was dynamically being formed, the question providing the furthest information regarding current estimated ability level of examinee was determined and it was presented to next examinee. Ability level of the examinee was re-estimated and updated according to the answer of each question. Thanks to this module, learners individually taught with UZWEBMAT are provided with adaptive assessment in accordance with their own qualifications and capacities. In contrast to classical tests, individual assessment becomes a reality through adaptive assessment which provides the opportunity to know learners according to their ability levels.